The XXI International Zhautykov Olympiad(IZhO) took place from January 13 to January 17 in Astana, Kazakhstan.
The results can be found here:
Scoreboard with points for each task is available in GDrive.
Problems can be found here:
Tutorials:
The tasks were authored and prepared by Tima, Aidos, bthero, sstrong, Mansur, DimmyT, Lokeo, dimachine, ArsenGotov, NaughtyMorzh, fractal.
We are looking forward to seeing you at IZhO 2026.
Is there a scoreboard that shows the breakdown per problem?
https://ranking.comprog.art/
Did anyone save the website while it was up? It seems to be down now, and I can't find a working version in web archive or similar sites.
Website is down now, I uploaded .csv ranking for both days, and .pdf ranking to GDrive.
Why does the text of the results link say 2020
To confuse participants I guess)
Does anybody have an idea why is this happening? (Why not perfect result?)
Sorry, problem was incorrectly uploaded.
All of the submissions for B are rejudged.
Problem B's big test cases seem to be weak. To test the intended solution's correctness, I submitted it with O(N^2) implementation, it got 78 :)
Same for me. C also has problems — selecting the part with less vertices in the bipartite graph to be the "maximal matching" passes on the tests.
Wow, that is interesting. Also, congratz for the win!
Thanks!
The tests for problem C1 have been fixed in the gym. Does it still pass the tests or was it during the main contest? By the way, sorry for wweak tests during the contest(
I know people have passed them during competition, haven't tried out the gym's ones yet.
I tried, but doesn't seem to pass. Also, what's the full solution for C1? I found a O(N^5) solution with Kuhn, also tried Dinic with O(N^4.5), but didn't passed last two subtasks.
The editorial will be published soon, but roughly you need to notice two ideas:
You only need to consider decreasing all values on a range (for $$$dp$$$ transition) to prefix minimum (larger is impossible, smaller is unnecessary).
If you fix left end of the range and move right end outwards, this minimum only decreases, thus, the graph constructed for current range will be a subgraph of the next range ($$$[l, r] \to [l, r + 1]$$$) and you can use current max matching as initialization for the next one (this dramatically speeds up Kuhn).
The first idea passes penultimate subtask, two ideas combined get full score. In total, it's supposed to be $$$O(N^{3.5})$$$.
The only 300 on the first day had both solutions for B and C wrong)
libobil orz!!!
when will editorials be posted?
We will publish editorials in a few days.
why cant i see others solutions?
fractal said he enabled viewing others' solutions, and it was working for a bit, but now it isn't working anymore for me. Is anyone else having this issue?
I can see solutions for a past contest, like this one:
but not the 2025 problems:
for problem B is the solution:
you do a dfs on the tree and maintain a segment tree (what is the maximum subpath(skeleton) ending at the current node U if the branch used for node U is of weight d) when going down to the child V you want to first remove the possibility of taking a brach with length w(U,V) if its occurence <= 2 then to build V's segment tree you want to query U's segment tree for each possible branch of node V of length d what is the maximum subpath ending at node U with branch length <= d
If not please share your solution)
editorial when?
How to solve Ostriches for 100? I think it has something to do with aliens trick. Because if the dp state is dp[pos][pics taken] = max distinct positions, then dp[pos][x] — dp[pos][x-1] >= dp[pos][x+1] — dp[pos][x] is true.